Draft framework for assessing residual capacity of
earthquake-damaged concrete buildings
K.J. Elwood and K. Marder
Dept of Civil and Env. Engineering, Univ of Auckland.
S. Pampanin, A. Cuevas Ramirez, M. Kral
Dept of Civil and Natural Resource Engineering, Univ of Canterbury
2016 NZSEE
Conference
P. Smith
SpencerHolmes Ltd., Wellington
A. Cattanach
Dunning Thornton Consultants Ltd., Wellington
M. Stannard
MBIE Building Performance Branch, Wellington
ABSTRACT: This paper presents a draft framework for assessing the residual capacity
of earthquake-damaged concrete buildings. Given the prevalent focus of past research on
the performance of undamaged buildings (new or existing), there is a dearth of
information on the performance of earthquake-damaged components and criteria for
assessment. This paper is the first step toward the development of a guideline for detailed
post-earthquake assessment of concrete buildings. The focus of this paper is the highlevel summary of a framework and the identification of future research needs to enable
the development of a reliable and practical residual capacity assessment procedure.
1 INTRODUCTION
The Canterbury earthquakes resulted in the demolition of over 60% of the concrete building stock in
the Christchurch Central Business District. While the factors influencing these demolition decisions
are complex and are integrally linked to the insurance held by the owner, in many cases a lack of
knowledge on the ability of earthquake-damaged RC buildings to resist future seismic loading was a
contributing factor in the decision to demolish (Marquis et al., 2015). This realization alerted the New
Zealand earthquake engineering community to the need for quantifiable methods that can reliably
assess the residual seismic capacity of earthquake-damaged concrete buildings.
Recognising this urgent need, MBIE Building Performance Branch established a Working Group in
2015 to develop consensus on practical means of assessing the residual capacity of reinforced concrete
buildings subjected to inelastic seismic demand. Considering ongoing changes in the insurance
market and the MBIE’s role in building regulations, the Working Group has focused on assessing the
safety of earthquake-damaged buildings; not on the assessment of buildings for the “as new”
reinstatement clauses in typical insurance policies in Christchurch.
This paper provides a draft framework proposed by the Working Group for the detailed assessment of
earthquake-damaged buildings. In developing this framework, it was recognised that considerable
research is required to make the framework usable; hence this paper serves the important function of
defining urgent research needs and will inform research supported by MBIE Building Performance
Branch and QuakeCoRE in the coming years.
The proposed framework is intended to be used by engineers engaged by a building owner for a
detailed assessment. This framework is not suitable for rapid assessments – e.g. placarding. The
objective is to assess the capacity of the building in terms of %NBS considering a future code-level
earthquake (return period of 500 yrs), regardless of the return period of the original earthquake, and
identify when repair of the building is feasible and warranted.
2 CHALLENGES FROM CHRISTCHURCH
Prior to the Canterbury Earthquake Sequence (CES), engineers undertook the structural assessments of
damaged buildings assuming material properties consistent with those used in the design of the
building on the basis that recent earthquakes in NZ rarely caused significant inelastic deformation of
reinforced concrete structures. Such assumptions were not considered appropriate after the CES in
view of the severity of shaking and evidence of plastic deformation demands in many buildings. The
following summarises concerns regarding the assessment of residual capacity, or reparability, of
damaged reinforced concrete buildings highlighted by experience in Christchurch.
Firstly, the extent of structural cracking in zones of plastic deformation varied significantly from that
which had been observed in laboratory cyclic load testing. During the 22 Feb 2011 earthquake, the
maximum demand occurred relatively early in the earthquake record; whereas under laboratory
conditions, components are typically subjected to progressively increasing inelastic deformations with
the maximum demands occurring at the end of the loading sequence. This typical approach to
laboratory testing allows extensive and interlinked cracking of the section to develop and potentially
greater yield penetration due to relaxation of bond with repeated cycles. Furthermore, the slow rate of
loading in typical laboratory tests can result in a great spread of cracking and plasticity compared with
components subjected to higher strain rates typical of earthquake demands. There was concern that
the limited number of cracks observed on concrete components in Christchurch buildings would
restrict the zones of inelastic deformations in the reinforcing steel, leading to high localised strain
demands and strain hardening in the reinforcement crossing the cracks. There was further concern that
strain aging of reinforcement may lead to embrittlement, further reducing the strain capacity in future
events (Erasmus and Pussegoda, 1977).
High tensile strength of concrete may have also contributed to a concentration of reinforcement strains
at a limited number of cracks. This feature of older structures with aged concrete, and those involving
precast concrete which utilised higher strength concrete for early handling and transportation,
increased the cracking strength relative to the yield moment, leading to the observed plastic hinges
with limited wide cracks (Henry, 2013). In particular, lightly reinforced shear walls performed poorly.
With high concrete strengths and low reinforcement ratios, in several cases the reinforcing steel
fractured and the gravity load resulted in the cracks closing, concealing the reinforcing steel failure.
In precast frame construction where construction joints were located within the plastic hinge zone,
virtually no concrete tensile capacity existed at the interface, and the initial crack occurred at the
construction joint with limited spread of plasticity beyond the single crack. The reinforcing steel at
such locations was assessed as likely to have been subjected to very large strain demands and
associated strain hardening. Again, combined with the influence of strain aging, the reduction in bar
strain capacity as a result of the plastic strain demands was unknown.
Another challenge was the lack of a clear relationship between the residual crack widths observed
after the earthquake and the maximum crack widths, and hence strain demands, that occurred during
the earthquake. Often damage to secondary elements, non-structural partitions and other brittle
elements provide more reliable evidence of the maximum displacements that occurred during the
earthquake shaking. A method of assessment is needed which provides a reliable estimate of the
maximum deformation demands during the earthquake.
Following the February 2011 earthquake it took a considerable period of time for the profession to
develop methods of assessing the level of strain hardening that was present in the reinforcing steel in
zones of inelastic deformation. A greater understanding of the characteristics of strain hardening and
strain aging has developed (Loporcaro et al., 2015), but further work is necessary to enable the
residual capacity of buildings damaged by significant earthquakes to be reliably established.
Further, the CES exposed buildings to multiple shallow aftershocks. The orientation, depth, and
severity of shaking varied between aftershocks and the analysis of buildings for this matrix of ground
shaking effects added significant complexity to the assessment of earthquake damage. The modelling
2
of buildings to reliably determine the impact of damage that occurred over several earthquakes
presented significant challenges to engineers undertaking building assessments. While the CES
resulted in numerous loading cycles on concrete components, many were below plastic limits and it
was not clear how many of these cycles must be considered and how the number of cycles affects the
residual capacity.
Other challenges to consultants undertaking the assessment of earthquake-damaged buildings were the
lack of documentation of the structural system, alterations to the building that were not well
documented, and accounting for foundation failures that limited the inelastic demand on the primary
structure or subjected the primary structure to differential settlements. Finally, guidance is needed on
effective methods of enhancing the deformation capacity of the plastic hinge zones where strain
hardening has potentially impacted the residual strain capacity (e.g. epoxy injection of cracks and/or
carbon fibre wrapping of the damaged zones).
In Christchurch, typical insurance policies which entitled the owner to a building “as when new” had
significant influence with respect to the decision to demolish or remediate buildings (Marquis et al.,
2015). In future earthquakes, a decrease in the affordability of insurance, or the rewording of
insurance policies, requires the profession to have the ability to assess the residual capacity of
earthquake-damaged buildings without undue conservatism. This paper proposes a framework for
assessing the residual capacity of reinforced concrete buildings damaged by earthquakes considering
many of the challenges faced in Christchurch.
Figure 1 - Proposed procedure for the detailed assessment of earthquake-damaged concrete buildings
3 PROPOSED FRAMEWORK
The framework described below, and summarised in Figure 1, provides the key steps for the detailed
assessment of earthquake-damaged concrete buildings. This is intended to be used for buildings
which exhibit damage in the seismic force resisting system consistent with good energy dissipation
3
and flexural hinges. The framework focuses on assessing the residual capacity of plastic hinges and
how this impacts the performance of the building as a whole (including the gravity system); no such
estimates are provided for members/systems exhibiting damage that is not consistent with flexural
hinging (e.g. shear failures). While the basic steps are described in as much detail as possible below,
gaps in knowledge have been identified and it is clear that research is urgently needed to make this
procedure usable by the engineering community. Research needs are identified in red italics.
3.1 Review building drawings
An important distinction between rapid and detailed building assessments is the access to all available
information about the building prior to assessment. Where available, the following information must
be reviewed to develop an in-depth understanding of the building structural characteristics:
- As-built structural drawings
- Alteration (retrofit) structural drawings (if applicable)
- Engineering calculations (if available)
Careful attention should be paid to identifying the seismic force resisting system (in each orthogonal
direction) and the gravity system. In buildings constructed before 1970s, a seismic force resisting
system may not be designated or apparent. Likely locations of plastic hinges in the seismic force
resisting system should be identified based on the structural system and building geometry. Potential
locations of damage in the gravity system must also be identified, with special attention paid to
support of precast floor systems, support for precast stairs, and damage to gravity columns or slabcolumn connections potentially exposed to large displacement demands.
3.2 Develop building model and perform analysis.
After the qualitative assessment of the building drawings described above, the engineer needs to
establish the expected performance of the building during the damaging earthquake. This requires
selection of the expected ground motion at the building site and development of a building model for
structural analysis. The engineer may elect to use linear or nonlinear analysis for this step; however,
an initial linear model is typically sufficient. Refer to AISPBE (NZSEE, 2016) for guidance on
selection of analysis procedures.
The characterisation of the ground motion (e.g. acceleration wave form or response spectrum)
experienced at the building site will depend on the analysis type selected. Unless doing nonlinear
dynamic analysis, typically a response spectrum will be sufficient. Ground motion at the building site
will typically not be available and judgement must be used to select an appropriate ground motion
considering both distance from building site to recording station and the site conditions of the building
site and recording station. If only a response spectrum is desired, response spectrum ordinates can be
determined based on a combination of spectral values from recorded stations and estimates from
Ground Motion Prediction Equations (Bradley and Hughes, 2012); however, attention should be paid
to capturing peaks and valleys in the response spectrum from recorded ground motions around the
period of the building (and at higher mode periods for tall buildings).
The primary objective of this initial analysis is to identify the expected building mechanism, confirm
locations of plastic hinges identified in step 1, estimate peak deformation demands (e.g. interstory
drifts and plastic hinge rotations), and identify any critical structural weaknesses. This analysis is
similar to analysis done for the purpose of pre-earthquake assessment of seismic capacity based on
AISPBE, except that the ULS demand spectrum is replaced by the response spectrum for the ground
motion at the building site. Any previous assessments of the building should be referenced in
developing this model and conclusions. Refer to AISPBE for guidance on building assessment
procedures.
3.3 Inspect building
Detailed inspection of the building is necessary to establish the building performance during the
damaging earthquake. The goal of inspection is to identify a range of damage measures from global
system performance to local strain demands. It is recognised that only residual damage can be
4
measured after the earthquake, and such measurements do not reflect the peak demands experienced
by the building during the earthquake. Key metrics are listed below as system, component, or
material-level measures.
System-level – The primary objective of the system-level inspection is to determine the primary
mechanism experienced by the building during the earthquake and the residual story displacements.
Residual displacements would typically be determined using a building survey, with special attention
paid to possible three-dimensional response of the building, i.e. if torsion resulted in one side of the
building experiencing larger residual displacements (damage) than the other. (While such response is
typically critical for torsionally irregular buildings (e.g. offset core), asymmetric response can still
develop in buildings with regular plan layouts if one seismic force resisting system experiences
yielding before another). The building survey will provide a profile of the building residual
displacements which can be used, in combination with the expected plastic hinge locations determined
in steps 1 and 2, to prioritise inspection of the structural system to identify plastic hinges. Based on
the location of plastic hinges identified in the structural system, the primary plastic mechanism
developed during the earthquake can be established. The engineer doing building inspection should
bear in mind that some of the observed damage may be the result of aftershocks with varying
orientation and intensity, emphasising the urgency of doing site inspection as quickly as possible after
the primary damaging earthquake.
Component-level – For each identified plastic hinge, crack maps should be sketched and crack widths
measured. Residual crack widths will depend on the axial load on the component and thus guidance
on the assessed hinge length will depend on the axial load. Assessed hinge length will be used to
estimate the average curvature and strain demands in plastic hinge.
Procedure needed to determine locations and length of plastic hinges based on axial loads and
measured crack widths. Need to define a total crack width along a specific axis of the component (e.g.
centre line) above which a plastic hinge is considered to have occurred. Length of hinge may be
determined based on a set distance (to account for strain penetration) beyond the last crack exceeding
a width of X mm. Need to confirm procedure with test data to ensure measured hinge provides good
relation between component displacements and measured average strains in hinge region. Note that
most existing plastic hinge models were developed to estimate ultimate drift capacity and hence may
not be appropriate at lower drift demands observed during a given earthquake. Furthermore existing
models are based on laboratory quasi-static reversed cyclic tests which may overestimate the extent of
cracking in the hinge region and thus overestimate the length of the plastic hinge. In the absence of
further research, the model by Priestley et al (2007) is recommended.
Material-level – Maximum crack widths, along with estimates of strain penetration (e.g. Priestley et
al., 2007), can be used to estimate residual strain in reinforcement. For low levels of damage,
measurement of crack widths should be sufficient to estimate residual bar strain; however, at high
damage levels more detailed material tests will be useful. Hardness testing procedures (both field and
lab) have been widely used following the Canterbury Earthquakes. Engineer is referred to Loporcaro
et al. (2015) for further discussion of hardness tests including limitations of application. Caution: for
lightly reinforced sections, fractured bars can be hidden at closed cracks in elements under axial load
(Henry, 2013).
Research is needed to identify criteria for initiating detailed material testing – i.e. maximum crack
widths above which it is recommended to perform more detailed material tests to confirm
reinforcement strain demands. Further research is necessary to determine if hardness tests provide a
reliable measure of peak strain demands after a complex cyclic history as expected due to earthquake
shaking.
3.4 Compare peak demands from model and observed damage distribution in building.
The objective of this step is to reconcile any differences between the results of from the structural
model (step 2) and field observations (step 3). Recall that the structural model from step 2 provided
peak demands, while observations in step 3 provided residual deformations, hence specific values
cannot be compared, but rather, focus on the distribution of demands throughout the building. Note
5
that residual displacement demands estimated from linear models are not meaningful and those from
nonlinear models are not typically reliable; hence, residual deformations should not be compared in
the reconciliation of the two procedures. Damage to secondary structure, such as yield line cracking
of floor slabs or damage to timber partitions, often provides a useful guide as to maximum
deformations during the damaging earthquake.
Research required to establish a relationship between observed residual damage and peak demands,
including collation of available experimental data on damage vs. drift relationship and development of
validation methods based on this data. Possibly require additional destructive or non-destructive tests
on damaged in-situ components or materials. Hardness tests, or equivalent, may be able to provide an
estimate of the peak strain demands in the reinforcement.
Differences between the model results and the field observations may point to field conditions not
considered in the modelling, e.g. soil-structure interaction or interaction with stiff non-structural
elements (NEHRP (2012) is a useful reference with respect to the consideration of soil-structure
interaction.). Model should be iteratively updated to achieve a better match with the observed damage
distribution from step 3.
3.5 Estimate demands from damaging earthquake
Based on the final model established in step 4 and the field observations in step 3, the following
demands from the damaging earthquake can be estimated:
 Demands from updated building model and estimated building site ground motion
o Peak Global (story) from displacement spectra and mechanism
o Peak Component curvatures (rotations) using measured hinge lengths
o Equivalent number of cycles (e.g. Hancock and Bommer, 2005)
 Demands from building inspection
o Residual Material demands (rebar strain) from crack widths and peak rebar strain from
hardness tests
o Residual Component curvatures (rotations) using measured hinge lengths
3.6 Assess if performance controlled by Low Cycle Fatigue (LCF)
Structural components can sustain a limited number of cycles prior to significant strength and stiffness
deterioration. The number of cycles to failure is dependent on the loading protocol, drift demands,
and failure mode (Hancock and Bommer, 2006). Equivalent number of cycles is used in the
assessment of liquefaction and slope stability and empirical equations for number of cycles have been
developed for this application (e.g. Stafford and Bommer, 2009). Design and assessment of structural
systems is done based only on peak demands, without consideration of number of cycles. If a prior
earthquake has consumed a portion of the cyclic capacity of the structural components, it may be
prudent to assess the residual capacity in terms of the number of cycles to LCF failure.
Research is needed to determine what characteristics of the damaging ground motion (e.g. pulse vs
long duration) will require consideration of LCF. This will identify when it is critical to include LCF
in the assessment of residual capacity.
Research is ongoing (Cuevas et al., 2015) to develop a procedure to determine the residual capacity in
terms of the number of cycles to LCF failure. This procedure involves the following steps:
 Determine the number of cycles to LCF failure at the peak drift demand from the damaging
earthquake.
 Estimate the effective number of cycles at the peak drift demand from the damaging earthquake.
 Calculate the consumed fraction of fatigue life as effective number of cycles from damaging
earthquake (demand) divided by the number of cycles to LCF failure (capacity).
6
3.7 Determine damaged (or repaired) component characteristics:
The objective of this step is to determine the stiffness, strength and deformation capacity of damaged
(or repaired) components based on peak and/or residual deformation demands from the damaging
earthquake established in step 5 (Epoxy injection of cracks is assumed as the form of repair). For
tension-controlled flexural components, strength degradation is expected to be minimal, while
stiffness degradation can be significant and will lead to larger drift demands in future earthquakes.
FEMA 306 (ATC, 1998) provides initial values for degradation of stiffness, strength and deformation
capacity for structural walls; however, such values do not include impact of limited plastic hinge
lengths, strain ageing of reinforcement, and number of cycles.
Research is ongoing to establish models for stiffness and strength degradation based on peak (and/or
residual) deformation demands established in step 5. Questions to be addressed in this research
include:
 Can the degradation models be based entirely on prior deformation demands or do they need
to include some measure of number of cycles or energy demand? What is error in estimate for
simpler approach?
 Can strength degradation be ignored for tension-controlled flexural components?
 How much stiffness degradation can be recovered with epoxy injection? Does this depend on
the crack widths from the damaging earthquake?
Deformation capacities of unrepaired damaged components may be reduced by embrittlement of
reinforcement due to strain ageing. For an unrepaired component, strain is expected to localise at the
same cracks developed in the damaging earthquake, hence reduced strain capacity of reinforcement
due to strain ageing must be accounted for.
Research is needed to confirm the above assumption that strain ageing will reduce the member
deformation capacity of an unrepaired component and identify the level of damage (e.g. crack widths)
at which strain ageing needs to be considered. Finally, peak and/or residual demands from step 5
need to be related to a reduction in strain capacity due to a combination of prior strain hardening and
strain ageing.
Repaired components by means of epoxy injection are expected to develop new cracks when subjected
to subsequent ground shaking. This is assumed to result in concentrated strain in reinforcement at a
different location than during the initial damaging ground motion, and consequently the impact of
strain aging may be ignored.
Research is needed to confirm the above assumption that strain aging can be ignored for repaired
components and establish the effectiveness of epoxy injection, particularly for construction joints in
the plastic hinge zone. The deformation capacity of the repaired component may depend on the
degree of observed damage (e.g. if spalling is observed, microcracking of concrete may reduce
deformation capacity even if strain aging can be ignored).
3.8 Update building model and conduct %NBS analysis:
The objective of this step is to assess the expected performance of the damaged or repaired building
for ULS demand (it is noted that aftershock risk should be assessed during the Level 2 assessment
stage (NZSEE, 2009)). This is done using the procedures of AISPBE and the stiffness, strengths, and
deformation capacities determined in step 7.
Two cases need to be considered:
a. Unrepaired damaged building subjected to design (ULS) earthquake demand: For this case the
objective is to ensure the overall performance of the building in a future earthquake has not
been significantly impacted by the observed damage. Special attention must be paid to the
drift demands on the gravity system considering the state of the gravity system after the initial
damaging earthquake determined in step 3 (e.g. remaining seating length for precast floor
units). The NZS 1170.5 design spectrum is used and %NBS for the damaged building is assessed according the procedures in AISPBE, using deformation capacities assessed in step 7.
7
If damaged %NBS < A, demolition of the building is recommended, while if A < damaged
%NBS < B, repair would be the appropriate action.
b. Repaired building subjected to design (ULS) earthquake demand: If repair is required based on
case (a), repaired component characteristics determined in step 7 are used to assess the overall
performance of the repaired building in a future earthquake in a manner similar to that for an
undamaged building using %NBS following AISPBE. The NZS 1170.5 design spectrum is
used for the demand. The repair will be considered acceptable if the %NBS for the repaired
building is greater than B.
Research and committee deliberations are needed to establish the criteria for repair (i.e. values of A
and B). Relationship of this assessment with earthquake-prone building policy also needs to be addressed.
4 SUMMARY OF RESEARCH NEEDS
Significant research needs have been identified in the development of this framework. A high-level
summary of the identified research needs is provided below.
 Procedure needed to determine locations and length of observed plastic hinges based on axial
loads and measured crack widths.
 Need to determine a maximum crack width beyond which detailed assessment of bar strain is
considered necessary.
 Establish a relationship between observed damage and peak demands experienced by
components.
 Determine what characteristics of the damaging ground motion (e.g. pulse vs long duration)
will require consideration of low-cycle fatigue (LCF).
 Establish a methodology for assessing the residual capacity in terms of the number of cycles to
LCF failure.
 Establish models for stiffness and strength degradation based on peak (and/or residual)
deformation demands. Determine how much stiffness degradation can be recovered with
epoxy injection of cracks.
 Establish when strain ageing needs to be considered in the residual capacity of damaged and
repaired components.
 Establish criteria based on the overall building capacity for determining when a repair or
demolition should be recommended.
Several of these research needs are currently being addressed by ongoing research programmes at the
Universities of Auckland and Canterbury, funded by the Natural Hazards Research Platform, MBIE
Building Performance Branch, and QuakeCoRE.
REFERENCES
Applied Technology Council (ATC) (1998) FEMA 306: Evaluation of earthquake damaged concrete and
masonry wall buildings: Basic Procedures Manual. Federal Emergency Management Agency, Washington.
Bradley, B. A., & Hughes, M. (2012). Conditional peak ground accelerations in the Canterbury earthquakes for
conventional liquefaction assessment. Technical Report for the Ministry of Business, Innovation and
Employment, New Zealand.
Cuevas, A., Pampanin, S., Carr, A., Ozbolt, J. (2015) Seismic residual capacity of reinforced concrete frames:
Part I: General Framework. Research Report UC 2015-4, Part I, Department of Civil and Natural Resources
Engineering, University of Canterbury, Christchurch, New Zealand.
Erasmus, L. and Pussegoda L. (1977) Strain age embrittlement of reinforcing steels. New Zealand Engineering
8
32(8).
Hancock, J. and Bommer, J.J. (2005) The effective number of cycles of earthquake ground motion. Earthquake
Engineering and Structural Dynamics: 2005, Vol. 34, pp. 637-664.
Hancock, J. and Bommer, J.J. (2006) A State-of-Knowledge Review of the Influence of Strong-Motion Duration
on Structural Damage. Earthquake Spectra: August 2006, Vol. 22, No. 3, pp. 827-845.
Henry, R. S. (2013). Assessment of minimum vertical reinforcement limits for RC walls, Bulletin of the New
Zealand Society for Earthquake Engineering 46, 88–96.
Loporcaro, G., Pampanin, S., & Kral, M. (2015). Experimental validation of “the hardness method” to estimate
the residual ductility of plastically deformed steel reinforcement. Proc., 2015 New Zealand Society of
Earthquake Engineering Conference.
Marquis, F., Kim, J. H., Elwood, K. J., & Chang, S. E. (2015) Understanding post-earthquake decisions on
multi-storey concrete buildings in Christchurch, New Zealand. Bulletin of Earthquake Engineering, DOI
10.1007/s10518-015-9772-8.
NEHRP (2012). Soil-structure interaction for building structures. NIST GCR, 12-917.
NZSEE (2009). Building safety evaluation during a state of emergency: Guidelines for Territorial Authorities.
New Zealand Society for Earthquake Engineering.
NZSEE (2016). Assessment and improvement of the structural performance of buildings in earthquakes (under
development). New Zealand Society for Earthquake Engineering.
Priestley MJN, Calvi GM and Kowalsky MJ (2007) Displacement-Based Seismic Design of Structures. IUSS
Press, Pavia, Italy.
Stafford, P.J., and Bommer, J.J. (2009) Empirical equations for the prediction of the equivalent number of cycles
of earthquake ground motion, Soil Dynamics and Earthquake Engineering, Volume 29, Issues 11–12,
November–December 2009, Pages 1425-1436, ISSN 0267-7261.
9